Best for reasoning

Claude Fable 5.1

Anthropic·1

271.7score

GPT-6 Astra

OpenAI·4

191.7score

Claude Opus 5.5

Anthropic·1

346.7score

Reasoning: ranks on every board

	Reasoning
Model · podium score	AA Intelligence	HLE	GPQA Diamond	Epoch ECI	ARC-AGI-2	LiveBench	MMLU-Pro
1GPT-6 Astra91.7	3					4	·
2Claude Fable 5.171.7	2	2	·	2	5		·
3Claude Opus 5.546.7		·	·	·	2	2	·
4Claude Fable 535.0	·	·	·	3	6	3	·
5Claude Opus 528.3	·	·	10	4	4	9	·
5GPT-5.6 Sol28.3	·	·	11	6	3	7	·
7Gemini 3.8 Flash26.7	15	4	2	13	·	·	·
8GPT-5.4 Pro25.0	·	5	4	9	12	·	·
9Gemini 3.1 Pro Preview23.3	·	3	5	·	13	22
10Muse Spark 1.321.7	4	·	·	15	·	5	·

#1 with the vendor’s logo23 podium7 top 10 stale board, not countedArrow keys move, Enter opens

GPT-6 AstraClaude Fable 5.1Claude Opus 5.5Claude Fable 5GPT-5.6 SolClaude Opus 5Claude Opus 4.5?gpt-5-1-codexGPT-5.5GPT-5.4

Artificial Analysis Intelligence Index

#1 Claude Opus 5.5, 57.6

Humanity's Last Exam

#1 GPT-6 Astra, 54.8%

GPQA Diamond (Epoch AI runs)

#1 GPT-6 Astra, 95.8%

Epoch Capabilities Index

#1 GPT-6 Astra, 166.6

ARC-AGI-2

#1 GPT-6 Astra, 95.0%

LiveBench

#1 Claude Fable 5.1, 83.4%

MMLU-Pro

#1 Gemini 3.1 Pro Preview, 91.2%

Stale since 2026-03-11

Hugging Face Open LLM Leaderboard

Retired

Retired

Scores as published by each board on the capture date. Model names and logos belong to their owners; logos via logo.dev.